pubs.acs.org/Langmuir © 2009 American Chemical Society
Facet-Selective Platinum Electrodeposition at Free-standing Polycrystalline Boron-Doped Diamond Films Ileana Gonzalez-Gonzalez,† Estev~ao Rosim Fachini,† M. Aulice Scibioh,† Donald A. Tryk,† Michele Tague,‡ Hector D. Abru~na,‡ and Carlos R. Cabrera*,† † Center for Advanced Nanoscale Materials and Department of Chemistry, University of Puerto Rico, Rı´o Piedras Campus, P.O. Box 23346, San Juan, Puerto Rico 00931-3346, and ‡Department of Chemistry and Chemical Biology, Cornell University, Ithaca, New York 14853-1301
Received October 21, 2008. Revised Manuscript Received June 9, 2009 In the present investigation, electrochemical deposition of platinum particles was carried out on boron doped diamond (BDD) films by using cyclic voltammetry at different potential sweep rates while maintaining the Pt concentration and number of potential cycles during the deposition as constant for all samples. The BDD film surfaces were studied using Raman spectroscopy, X-ray diffraction, and scanning electrochemical microscopy. The deposited particles were characterized by scanning electron microscopy/X-ray energy dispersive analysis, X-ray photoelectron spectroscopy, and cyclic voltammetry before and after methanol oxidation. The platinum nanoparticles are found to be selectively electrodeposited on the (111) facets of the BDD. In addition, the location of the Pt particles on the diamond facets was affected by the potential sweep rate. For higher sweep rates, the particle size was dependent on the facet on which the particles are electrodeposited with smooth (110) facets exhibiting a smaller number of particles but with a larger particle diameter. After methanol oxidation studies using cyclic voltammetry and controlled potential electrolysis for several hours, the platinum particles remained attached to the (111) facets of the BDD, while the particles on the (110) facets of the BDD became agglomerated along grain boundaries. Functional groups present on the (111) facet of the diamond surface play an important role on the stability of the particles attached to the diamond surface. After methanol oxidation, the particles deposited on other facets appeared to lose their adhesion leading to agglomeration on the grain boundaries. BDD appears to be a promising electrocatalyst support material that can help to resist platinum nanoparticle agglomeration in direct methanol and other low temperature fuel cell applications.
1. Introduction Metal nanoparticle catalysts on electrode surfaces can show high catalytic performance, and the activity is a function of particle size and the nature of the support as well as the preparation route employed.1 For the support, a conductive material such as carbon black, with sp2-bonded carbon, is typically used in fuel cell systems.2 This support is also a high surface area material used to physically separate the catalyst particles and decrease their agglomeration when in use. The high surface-to-volume ratio of the supported platinum particles in these electrocatalysts maximizes the active area available for reaction. Polymer electrolyte fuel cells (PEMFCs) can be operated using low platinum loadings provided the platinum particles are appropriately dispersed throughout the carbon matrix. If the distribution/dispersion of the platinum nanoparticles cannot be maintained over the lifetime of the fuel cell, the electrochemical activity of the cell will deteriorate over time. The conventional carbon blacks that are used as catalyst support are susceptible to microstructural and morphological degradation under the oxidizing conditions existing at the oxygen *Corresponding author. E-mail:
[email protected]. Phone: 787 764 0000 x14807#. (1) Zhang, Y.; Asahina, S.; Yoshihara, S.; Shirakashi, T. Electrochim. Acta 2003, 48, 741. (2) Paik, C. H.; Jarvi, T. D.; O’Grady, W. E. Electrochem. Solid-State Lett. 2004, 7, A82. (3) Shanna, D.; Knights, K.; Colbow, M.; St-Pierre, J.; Wilkinson, D. P. J. Power Sources 2004, 127, 127. (4) Wang, J.; Swain, G. M. Electrochem. Solid-State Lett. 2002, 5, E4. (5) Hupert, M.; Muck, A.; Wang, J.; Stotter, J.; Cvackova, Z.; Haymond, S.; Show, Y.; Swain, G. M. Diamond Relat. Mater. 2003, 12, 1940.
Langmuir 2009, 25(17), 10329–10336
cathode.3-5 Degradation of the electrocatalyst support is a serious problem because it leads to activity loss due to catalyst detachment, agglomeration, and a general mechanical failure of the electrodes. This effect also increases the ohmic resistance and degrades the operational efficiency of PEMFCs.2,6-8 Thus, the development of advanced support materials that are stable at high potentials, low pH, and relatively elevated temperatures is still the subject of much study.9 The activity and corrosion resistance of such electrodes has been found to depend markedly on the carbon substrate used. Electrodes can be prepared with films of more corrosion-resistant particles such as graphitized carbon blacks,10 carbon nanotubes,11,12 carbon fibers,13 and boron-doped diamond (BDD).14,15 The sp3-type bonding of the diamond surface makes it less reactive than sp2bonded carbon black. (6) Kangasniemi, K. H.; Condit, D. A.; Jarvi, T. D. J. Electrochem. Soc. 2004, 151, E125. (7) Wang, J.; Swain, G. M.; Tachibana, T.; Kobashi, K. Electrochem. SolidState Lett. 2000, 3, 286. (8) Antolini, E. J. Mater. Sci. 2003, 38, 299. (9) Huang, H. X.; Chen, S. X.; Yuan, C. J. Power Sources 2008, 175, 166. (10) Yang, Y.; Zhou, Y.; Cha, C. Electrochim. Acta 1995, 40, 2579. (11) Tang, H.; Chen, J. H.; Huang, Z. P.; Wang, D. Z.; Ren, Z. F.; Nie, L. H.; Kuang, Y. F.; Yao, S. Z. Carbon 2004, 42, 191. (12) Han, K. I.; Lee, J. S.; Park, S. O.; Lee, S. W.; Park, Y. W.; Kim, H. Electrochim. Acta 2004, 50, 791. (13) Mahmood, T.; Williams, J. O.; Miles, R.; McNicol, B. D. J. Catal. 1981, 72, 218. (14) Salazar-Banda, G. R.; Eguiluz, K. I. B.; Avaca, L. A. Electrochem. Commun. 2007, 9, 59. (15) Gonzalez-Gonzalez, I.; Tryk, D. A.; Cabrera, C. R. Diamond Relat. Mater. 2006, 15, 275.
Published on Web 07/27/2009
DOI: 10.1021/la8035055
10329
Article
Gonz alez-Gonz alez et al.
BBD films exhibit very high stability to vigorous chemical and electrochemical treatments.2,16 Diamond films prepared by chemical vapor deposition have been employed in numerous applications due to their unique characteristics. Various studies on metal particle deposition on BDD have been carried out to examine the catalytic properties of the films.17 Thus far, there have been few studies on the electrodeposition of metal nanoparticles on diamond and on the selective adsorption of metal nanoparticles on specific facets (ref 18 and the references cited therein). Because boron is known to incorporate into the (111) growth sectors more abundantly (up to a factor of a 100) than into (100) growth sectors,19,20 we have set to examine the possibility that electrodeposition might occur preferentially on the (111) facets as well as on defect sites where boron tends to incorporate to a greater extent during deposition. In the present work, we have used noble metal electrocatalyst particles deposited on BDD films as model systems for high-area electrocatalyst supports to characterize the interaction of platinum on the diamond surface.
2. Methodology A standard three-electrode assembly was employed in cyclic voltammetry for the deposition of platinum particles on the BDD films as well as in the methanol oxidation studies. The BDD films were used as working electrodes, a Ag/AgCl electrode, and a Pt wire were used as the reference and counter electrodes, respectively. BDD films were bought from Element Six, 10 mm 10 mm 1 mm thickness, surface finish grown by chemical vapor deposition with a resistivity between 0.038 and 0.105 Ω-cm, and a doping level of [B] > 1020 cm-3. Each diamond film was cleaned by using a solution of aqua regia, followed by cycling from -0.8 V to þ1.6 vs Ag/AgCl in sulfuric acid 0.5 M, and electrochemically tested as clean sample of boron doped diamond. The secondary ion mass spectrometry (SIMS) analyses were performed on as-received polycrystalline diamond samples in a PHI 5600 Multisystem equipped with a differential pressure ion gun and a quadrupole mass analyzer. The primary ion was Arþ at 4 kV and 45° of incident angle. Charging compensation was used despite the quasimetallic characteristics of the sample because it contributes for the accuracy of analysis. The beam density current was previously calibrated to 10 nA/mm2 using a Faraday cup. An electronic gate of 70% was selected, with consequent reduction of the area of analysis compared to the sputtered surface in order to avoid interferences from regions outside of the interest area. The base pressure of the system was established at 2 10-8 mbar and then raised up to 2 10-7mbar introducing oxygen gas. The oxygen leak was used to improve the counts of mass peaks. Those conditions correspond a nonstatic SIMS operation, and they were selected in order to obtain information from the near outermost volume of the sample (static SIMS conditions just probe a very few amount of the most external layer, which could mislead the real amount of B on the diamond outermost volume). After all conditions were tested in a previous sample, the spectra recorded the first two scans of an untouched sample surface in the range of 0-50 uma. The time of analysis was selected by considering reproducible spectra; after that time, the spectra will change, probably due to different composition and charge effects when deeper matrix regions are probed. Graphite (HOPG) and natural diamond spectra were also recorded under the same conditions (16) Pleskov, Y. V.; Evstefeeva, Y. E.; Krotova, M. D.; Varmin, V. P.; Teremetskaya, I. G. J. Appl. Electrochem. 2003, 33, 909. (17) Gao, J.; Arunagiri, T.; Chen, J.; Goodwill, P.; Chyan, O.; Perez, J.; Golden, D Chem. Mater. 2000, 12, 3495. (18) Kondo, T.; Aoshima, S.; Hirata, K.; Honda, K.; Einaga, Y.; Fujishima, A.; Kawai, T. Langmuir 2008, 24, 7545. (19) Locker, R.; Wagner, J.; Fucks, F.; Maier, M.; Gonon, P.; Koidi, J. Diamond Relat. Mater. 1995, 4, 678. (20) Kanda, H.; Sekine, T. Properties and Growth of Diamond, Davies, G., Ed.; IEE: London, 1994.
10330 DOI: 10.1021/la8035055
for comparison purposes. Because the present study is related to the surface, no depth profile of B implantation was investigated. Fisher Optima grade sulfuric acid solution, diluted to 0.5 M with ultrapure water, was used as the supporting electrolyte. All measurements were done with a BAS Epsilon potentiostat from Bioanalytical Systems, Inc. A specially designed electrochemical cell was used to conveniently fit the diamond film with an O-ring seal. The Pt electrodepositions were done using solutions made with K2PtCl6 3 6H2O Aldrich reagent grade (98%). Electrochemical depositions were done by cyclic voltammetry. The potential was scanned between -0.2 and 1.0 V vs Ag/AgCl in 1 mM K2PtCl6/ 0.5 M H2SO4 solution. The potential scan rate was varied for different sample preparations: 100, 250, and 500 mV s-1. To study the catalytic performance of the deposited particles, the methanol oxidation onset was measured using cyclic voltammetry and controlled potential chronoamperometry. The potential was cycled between -0.2 and 1.0 V vs Ag/AgCl in a solution of 0.5 M CH3OH/0.5 M H2SO4, and then we held the potential constant at 0.4 V vs Ag/ AgCl for 1800 s (Potentiostat Limited). The various electrodeposited catalysts were characterized by scanning electron microscopy-X-ray energy-dispersive analysis (SEM/EDS), X-ray diffraction, and cyclic voltammetry. For the SEM micrographs, a JEOL 5800LV microscope was used at 15 kV. X-ray diffraction analyses were performed using a SIEMENS D5000 X-ray diffractometer using a Cu KR polychromatic X-ray source. The Raman spectra were obtained with an ISA-JY T64000 spectrometer. The measurements were performed in the backscattering configuration on a spot of about 3 mm in diameter illuminated with an Arþ laser (514.5 nm, 10 mW). Scanning electrochemical microcopy measurements were done using a CHI900 SECM with a Pt ultramicroelectrode with a diameter of 10 μm. The solutions were made in 0.5 M H2SO4 solution. The electrochemical cell was completed using a Pt wire and Ag/AgCl as counter and reference electrodes, respectively. Platinum particle size distribution measurements were performed using the UTHSCSA Image Tool program21 in which the SEM images were imported and the sizes of the white pixels were measured. This was done in the manual mode to avoid incorrect identification of the particles.
3. Results and Discussion Three free-standing diamond films with the same boron doping level were investigated using X-ray diffraction (XRD) analysis (figure not shown), and all of them exhibited a similar diffraction pattern consisting of three peaks, which are attributable to the diamond structure. The first peak appeared at 43.84°, due to (111) plane, and a second peak, at 75.21°, was characteristic of the (220) plane. The (220) peak was more intense than the (111) peak; this is because the growth process occurred preferentially in the (220) plane. The third peak, at 91.38°, was assigned to the (311) plane. To find out the quality of the diamond films employed in this study, Raman spectral analyses were done. From the Raman spectra (Figure 1), the peak at 1332 cm-1 confirmed the presence of sp3 carbons, as expected for diamond materials.22 A small peak in the 1598 cm-1 region indicated the presence of sp2 carbon impurities. The films were of good quality, although sp2 carbon contamination was present in trace amounts. Further, the freestanding boron-doped diamond films were subjected to X-ray photoelectron spectroscopic (XPS) analysis in order to determine the extent of boron doping levels in these diamond films. The carbon 1s binding energy region showed a strong peak corresponding to C-C bonds and two additional peaks assigned to C-O and CdO; this functionalization occurred (21) Developed at the University of Texas Health Science Center at San Antonio, Texas and available from Internet by anonymus FTP from Maxrad6. uthscsa.edu. (22) Knight, D.; White, W. J. Mater. Res. 1989, 4, 385.
Langmuir 2009, 25(17), 10329–10336
Gonz alez-Gonz alez et al.
Figure 1. Typical Raman spectra of the boron-doped diamond films used in the Pt electrodeposition study.
during the cleaning process of the diamond before Pt electrodeposition. XPS analysis revealed that all three freestanding boron-doped diamond films used in this work had a boron/ carbon atomic ratio of 0.03 on the diamond surface. The boron is expected to be incorporated mostly into the facet (111) and grain boundaries, as reported by Re et al.23 The SIMS analysis shows that the boron doped diamond film sample exhibited two peaks at 10 and 11 uma, both corresponding to Bþ species. These peaks are absent in the graphite and pristine diamond samples. For boron doped diamond films, the isotopic ratio 10B/11B can be calculated from those two peaks and it was 0.24, which is in good agreement to the natural isotopic ratio for that element (0.20)24 within experimental errors. It is difficult to quantify the element concentrations from SIMS spectra without proper calibration standards. Also, extrapolation of data for different matrix or ion sources must be done carefully in order to obtain accurate results. On the other hand, the diamond film supplier reported the doping level higher than 1.0 1020 B atoms cm-3.25 An additional difficulty is that SIMS does not give accurate quantification for heavily doped boron diamonds due to a change in the ionization probability for 12Cþ peak.26 Also, the boron doping is expected to be quite heterogeneous among different facets of diamond crystals. Kolber et al. reported values of 0.3-1.45% of B in diamond matrix and proved that the amount of B can be improved as high as by a factor of 5, changing from {100} to {111} diamond planes.27 Considering these issues, a tentative methodology has been adopted to estimate the concentration of boron in the sample. In the absence of standards, using relative selective factor (RSF) for diamond matrix24 (despite our oxygen leak conditions, the RSF used came from a different ion source, O2þ) and isotopic corrections, results in concentration of boron of 3.5 1020 atoms cm-3 (or 0.2%) were obtained, which agrees with the range of expected values. The reported value of boron doping was an average on a large area and does not distinguish for different diamond crystalline orientations. (23) Ri, S.; Kato, H.; Ogura, M.; Watanabe, H.; Makino, T.; Yamasaki, S.; Okushi, H. Diamond Relat. Mater. 2005, 14, 1964. (24) Wilson, R. G.; Stevie, F. A.; Magee, C. W. Secondary Ion Mass Spectrometry: A Practical Handbook for Depth Profiling and Bulk Impurity Analysis; John Wiley & Sons: New York, 1989; Table B2, p App B.5. (25) Element Six 2001-2009, British Isles (26) Guzman de la Mata, B.; Sanz-Hervas, A; Dowsett, M. G.; Schwitters, M.; Twitchen, D. Diamond Relat. Mater. 2007, 16, 809. (27) Kolber, T; Piplits, K; Haubner, R; Hutter-Fresenius, H. J. Anal. Chem. 1999, 365, 636.
Langmuir 2009, 25(17), 10329–10336
Article
Density functional theory (DFT) method has shown this behavior previously. The study found the adhesion effect at (111) surface and for BH* in a specific position on the (100) surface. The BH* adsorbate showed a particular capability to capture an H atom from the coadsorbate CH3, followed by formation of a new bond between the resulting coadsorbed CH2* and BH2 species.28 Two different types of regions on the diamond surface have been previously evidenced with metallic or semiconducting properties that were characterized with different types of Raman29,30 spectra and CPAFM.31 Scanning electrochemical microscopy (SECM) is an electrochemical scanning probe technique in which the measured current is due to an electrochemical reaction at the tip.32 The SECM can be used to map zones of gradually varying reactivity on a diamond electrode surface.33 By measuring the feedback current at intermediate potentials, it can be seen that the current increases in areas where the electron transfer kinetics are higher. At higher potentials, the rate of oxidation is diffusion-controlled, so the image shows a uniform reaction rate across the electrode surface in feedback mode. In Figure 2, a SECM feedback image of the diamond surface morphology is shown. This image was obtained by measuring the current when applying a constant potential of 0.8 V vs Ag/ AgCl to the tip of a Pt ultramicroelectrode, while the BDD film was not connected to the cell. The image in Figure 2 shows the current difference between the SECM image of the same area when no potential was applied to the BDD sample (top) and when -0.5 V was applied to the BDD sample in order to reduce the oxygen that is being produced at the platinum electrode (bottom). The SECM image shows that the oxygen reduction is not homogeneous throughout the BDD film surface. This result agrees well with those published previously by Macpherson et al.34 on their SECM using the reduction of Ru(NH3)63þ and with conducting atomic force microscopy measurements that determined two characteristic conductivity domains with resistances of 100 kΩ and 50 MΩ. Honda et al.35 used electrogenerated chemiluminescence (ECL) to map the spatial variations in electrochemical activity at the heavily doped polycrystalline diamond surface as well. The ECL intensities for (100)-oriented growth sectors were significantly lower than those for the other growth sectors and remained at ca. 50% of those for (111) sectors even at potentials at which the intensity reached a maximum. The electrochemical deposition of the platinum particles was done by cyclic voltammetry by varying the potential sweep rate for different depositions. We expected that at higher potential sweeps the electron transfer kinetics within different facets of the diamond are going to dictate the deposition, this difference becoming accetuated on thick diamond films like the ones employed in this study (1 mm). The platinum concentration (1.0 mM K2PtCl6/0.5 M H2SO4) and the number of potential cycles for the Pt electrodeposition were the same for all the (28) Regemorter, T. V.; Larsson, K. J. Phys. Chem. A 2008, 112, 5429. (29) Szunerits, S.; Mermoux, M.; Crisci, A.; Marcus, B.; Bouvier, P.; Delabouglise, D.; Petit, J.; Janel, S.; Boukherroub, R.; Tay, L. J. Phys. Chem. B 2006, 110, 23888. (30) Ushizawa, K.; Watanabe, K.; Ando, T.; Sakaguchi, I.; Nishitani-Gamo, M.; Sato, Y.; Kanda, H. Diamond Relat. Mater. 1998, 7, 1719. (31) Stotter, J.; Show, Y.; Wang, S.; Swain, G. Chem. Mater. 2005, 17(19), 4880. (32) Bard, A. J.; Fan, F. F. R; Mirkin, M. V. J. Electroanal. Chem. 1996, 68, 177A. (33) Lu, G.; Cooper, J. S.; McGinn, P. J. Electrochim. Acta 2007, 52, 5172–5181. (34) Wilson, N. R; Clewes, S. L.; Newton, M. E.; Unwin, P. R.; Macpherson, J. V. J. Phys. Chem B 2006, 110, 5639. (35) Honda, K.; Noda, T.; Yoshimura, M.; Nakagawa, K.; Fusjishima, A. J. Phys. Chem B 2004, 108, 16117.
DOI: 10.1021/la8035055
10331
Article
Gonz alez-Gonz alez et al.
Figure 2. (upper) Scanning electrochemical microscopy of a freestanding boron-doped diamond in a 0.5 M H2SO4 solution with a Pt tip potential 800 mV vs Ag/AgCl and a Pt wire as counter electrode. (lower) Comparison of scanning electrochemical microscopy of a freestanding boron-doped diamond in a 0.5 M H2SO4 solution with a Pt tip potential 800 mV vs Ag/AgCl and a Pt wire as counter electrode and -500 mV vs Ag/AgCl potential applied to BDD film.
samples. This deposition is based on the PtCl62- redox reaction.36 PtCl26 - þ 4e - TPt þ 6Cl - E 0 ¼ 964 mV vs Ag=AgCl The morphologies of the platinum nanoparticles electrodeposited by cyclic voltammetry were determined from scanning electron micrographs, and their composition was examined by using energy dispersive X-ray spectra (EDX) and X-ray photoelectron spectroscopy (XPS). Cyclic voltammograms of the platinum particle-decorated BDD film were obtained after cleaning cycles were completed; it showed the characteristic current features for platinum, hydrogen adsorption, and desorption and Pt oxide formation and Pt oxide reduction.37 The electrochemically active surface area of the electrode was estimated by integrating the anodic current peak obtained during the forward sweep (-0.2 to 0.1 V vs Ag/AgCl).38 Samples A, B, and C were deposited by cyclic voltammetry at scan rates of 100 mV s-1, 250 mV s-1, and 500 mV s-1, respectively. The platinum surface areas of the particles deposited on samples A, B, and C were 1.80 cm2, 6.14 cm2, and 7.02 cm2, respectively. The higher surface area of sample C may be due to the fact that, at higher sweep rates, more nucleation sites are formed on the diamond surface, and due to the rapid sweeps during the deposition, the particles appear to experience a slower growth rate. This may be due to the amount of time that the nuclei are held at the depositing potential (-0.2 V vs Ag/AgCl) by assuming a growth diffusion dependent process.39
(36) A. J. Bard and Faulkner, L. R.; , Electrochemical Methods. Fundamentals and Applications; John Wiley & Sons, Inc: New York, 2001; p 700. (37) Pattabiraman, R. Indian J. Chem. Technol. 1996, 3, 269. (38) Feltham, A. M.; Spiro, M. Chem. Rev. 1971, 71, 177. (39) Hyde, M. E.; Jacobs, R.M. J.; Compton, R. G. J. Electroanal. Chem. 2004, 562 61.
10332 DOI: 10.1021/la8035055
The high-resolution X-ray photoelectron spectrum (Figure 3A) shows the binding energy peak of Pt0 4f 5/2 and Pt0 4f 7/2, and with an increase of þ1.2 eV in the binding energy, the signal for platinum dioxide. The C 1s region (Figure 3A) has also been fitted to three binding energy peaks. One main peak at 284.5 eV and two additional peaks at þ1.4 and þ3 eV higher binding energies are observed. These peaks have been attributed to C-C, C-O, and CdO on BDD, respectively. The peak at þ1.4 eV, attributed to C-O, has been correlated40,41 to C-OH from (111) facets and C-O-C from (110) facets. The peak þ3 eV can be attributed to CdO from the (110) facet, even though the chemical shift is slightly smaller than that reported in the literature.42 The highresolution X-ray photoelectron spectrum (Figure 3B) shows the binding energy peaks of Pt0 4f5/2 and Pt0 4f7/2, and with an increase of þ1.2 eV in the binding energy, the signal for platinum dioxide is observed. The functional group O-H on facet (111) appears to be the nucleation site for the Pt particles on all the samples. As shown in Table 1, for XPS analysis on the carbon 1s binding energy region, the C-O/CdO peak area ratio decreases in each of the samples after the platinum deposition is done. On the oxygen region, the O-H/metal oxides ratio also decreased. These results show that the hydroxy groups on the diamond facet (111) are involved on the platinum particle nucleation on the diamond surface. Scanning electron micrographs (SEM) of samples A, B, and C before methanol oxidation studies are presented in Figure 4. As can be seen in this figure, sample A showed that the (40) John, P.; Polwart, N.; Troupe, C. E.; Wilson, J. I. B. J. Am. Chem. Soc. 2003, 125, 6600. (41) Wagner, C. D.; Naumkin, A. V.; Kraut-Vass, A.; Allison J. W.; Powell, C. J.; Rumble, J. R. NIST X-Ray Photoelectron Spectroscopy Database, NIST Standard Reference Database 20, version 3.4 (web version); U.S. Department of Commerce. (42) Kondo, T; Honda, K.; Tryk, D. A.; Fujishima, A. J. Electrochem. Soc. 2005, 152, E18.
Langmuir 2009, 25(17), 10329–10336
Gonz alez-Gonz alez et al.
Article
Figure 3. High resolution X-ray photoelectron spectra of the Pt particle decorated boron-doped diamond film. (A) C 1s, (B) Pt 4f, and (C) O 1s binding energy regions.
platinum particles are evenly distributed on the facets of the diamond surface, but at the same time, these particles are agglomerated. In sample B, the platinum particle deposition occurred exclusively on the (111) facets of BDD, and the rest of the facets remained clean. Sample C exhibited a selective deposition in which the particles were only present on one of the facets, which from the literature35 can be presumed to be the (111) facet. We concluded that this particular facet was (111) on the basis of the edge lines of the facet with an intersection angle of 60° and 120°, as can be clearly seen in Figure 5, as well as other micrographs not presented here. In the case of sample C, the particle distribution depended on the facet of the diamond surface. It can be seen that a greater number of smaller particles were deposited on one facet, and larger particles (agglomerated particles) were deposited on the other facets. The elemental chemical composition of the particles was verified by carrying out energy dispersive X-ray (EDX) analysis. The X-ray fluorescence signal for carbon, oxygen, and platinum can also be seen in Figure 5. In this study, the platinum particle location was found to be affected by the potential sweep rate used during the electrodeposition. At higher scan rates, in which the platinum deposition was selective to one facet of the BDD film, as seen in sample C. At a sweep rate of 250 mV s-1 (sample B), the deposition occurred only on the (111) facets and on defects. For the higher sweep rate, the particle size was dependent on the facet that was electrodeposited; smooth facets showed fewer particles with larger diameter. These observations suggest that the nucleation process is much faster than the particle growth and that the hydroxy (C-O-H) functional groups present on the (111) facet after anodic oxidation of the diamond surface promote the nucleation of the platinum particles. Platinum electrodeposition on BDD free-standing films, from Element6 and used as received, showed smaller Pt particles. We cannot comment on H-terminated films because XPS analysis done on these films showed a peak on the oxygen binding energy region. Freestanding films, without the anodic pretreatment, showed fewer deposited particles and thus a smaller particle distribution. In general, we obtained different particle size distributions for different facets of the BDD surface, depending on the rate of Langmuir 2009, 25(17), 10329–10336
the deposition, even though our particle distributions were of the same magnitude as those reported in the literature.14,15 The platinum particle size range and distribution were between 12 and 198 nm and 108 and 109 particles/cm2, respectively. For a slow scan rate, i.e., sample A, the particles were equally deposited and distributed through the treated diamond film. Nonetheless, particle agglomeration can be seen in the SEM images of the film. In the case of particles deposited at higher sweep rate, the distributions varied. For sample B, deposited at 250 mV s-1, the particle distribution was 2.8 108 particles/cm2 on the more conductive facet, but essentially no particles were present on other ones. It is important to mention that the distribution on this sample was lower than that for sample A, but this sample exhibited completely clean facets. Sample C exhibited different size distributions for different facets. For more conductive facets, the distribution was 8 109 particles/cm2, while for another one, the distribution was 1 107 particle/cm2. A geometrical particle analysis was done for each particle, and the values were used to construct the histograms presented in Figure 6. For sample A, the average particle ferret diameter was 75 ( 36 nm with a modal of 50 nm; for sample B, the average ferret diameter was 198 ( 89 nm with a modal of 275 nm, and sample C exhibited an average of 12 ( 9 nm with a modal of 9 nm. The average sizes of these particles are smaller than those for any previously reported Pt particles deposited on diamond in the literature.4,43 Metal particle electrodeposition on diamond films has been extensively reported by a number of groups. For example, Swain7 and Montilla43 obtained interesting, yet different, results from the ones presented here. Nanometer-size metal particles with preferred crystalline textures have been spontaneously deposited on diamond thin films after a simple immersion in an acidic solution containing metal ions or metal complex ions. The diamond/silicon interfacial ohmic contact was found to be the critical factor for achieving the observed spontaneous metal deposition on the diamond surface; this type of deposition was not observed on freestanding diamond films.17 (43) Montilla, F.; Morallon, E.; Duo, I.; Comninellis, Ch.; Vazquez, J. L. Electrochim. Acta 2003, 48, 3891.
DOI: 10.1021/la8035055
10333
Article
Gonz alez-Gonz alez et al.
Figure 5. Scanning electron microscopy image of sample (C) showing edge lines of the facets (111) with intersection angles 60° and 120° (upper) and the corresponding energy X-ray dispersive spectrum of platinum particle on BDD film (lower).
Figure 4. Scanning electron microscopy images of polycrystalline free-standing BDD films after Pt deposition by cyclic voltammetry in a 1 mM K2PtCl6/0.5 M H2SO4 solution at different scan rates: (A) 100 mV/s, (B) 250 mV/s, and (C) 500 mV/s. The potential was scanned between -0.2 and 1.0 V vs Ag/AgCl. Table 1 carbon region, 1s CO/CdO
BDD Pt-BDD
oxygen region, 1s (O-H)/metal oxides
(A)
(B)
(C)
2.4 1.40
4.59 2.67
5.82 2.81
(A) 6.05 1.39
(B) 3.64 2.57
(C) 16.48 6.68
Wang and Swain44 reported on the fabrication of platinum/ diamond composite electrodes. With their approach, the Pt (44) Wang, J; Swain, G. M. J. Electrochem. Soc. 2003, 150, E24.
10334 DOI: 10.1021/la8035055
particles were galvanostatically electrodeposited and, later, a diamond thin film was deposited on top of the Pt particles in order to physically secure them. The control of the particle size (30-500 nm) and distribution (∼108 to 109 particles/cm2) was achieved by varying the electrodeposition and secondary diamond growth conditions. R. Compton has also published on the growth mechanism of Au particle on diamond deposited by cyclic voltammetry and galvanostatically on BDD films bought also from Element6.39 The growth of our particles agrees with the nucleation-diffusion dependent growth. One possible explanation of why Compton et al. did not observe Au selective facet deposition is because they did not preoxidized the BDD film samples before the electrodeposition. Cyclic voltammetric and chronoamperometric measurements for methanol oxidation were performed on all three samples. Cyclic voltammetry was done at 24 °C and °65 C, and all the samples showed both peaks corresponding to methanol oxidation, with onset potentials range of (measured at 1 μA cm-2) 413449 mV (Figure 6 shows this measurements for 24 °C). Figure 7 shows a time vs current plot for the platinum particles in 0.5 M H2SO4 solution with 1 M CH3OH at 24 °C at a constant potential of 0.500 V vs Ag/AgCl for 1000 s, although similar results were obtained for 64 °C. As can be seen in the Figure 8, during methanol oxidation chronoamperometry, even though sample B showed higher currents at 25 °C, it exhibited a rapid decay. Thus, the current stability of sample C at both temperatures was better. There are two factors affecting the methanol oxidation Langmuir 2009, 25(17), 10329–10336
Gonz alez-Gonz alez et al.
Article
Figure 6. Scanning electron microscopy image of polycrystalline free-standing BDD films after Pt deposition by cyclic voltammetry in a 1 mM K2PtCl6/0.5 M H2SO4 solution by cyclic voltammetry at (A) 100 mV/s, (B) 250 mV/s, and (C) 500 mV/s (upper). Histogram of the ferret diameter of the electrodeposited Pt particles (lower).
Figure 7. Cyclic voltammetry on 0.5 M CH3OH/0.5 M H2SO4 Pt particles deposited on boron-doped diamond films by cyclic voltammetry at different scan rates: sample A at 100 mV/s ( 3 3 3 ), sample B at 250 mV/s (;), and sample C at 500 mV/s (---).
stability: (1) poisoning of the Pt particles, as established in the literature45 and (2) the agglomeration of the particles that were deposited on the less conductive facets. The current gradually decayed with time, and the initial currents for each case were smaller than those from voltammograms at the same potential. The decay in oxidation current was different for each sample and each temperature. The stability of these particles was excellent, as evidenced by their ability to resist strong acid conditions and extended electrolysis. It is important to mention that each of the samples was subjected to 30 cycles of methanol oxidation and 3 h of controlled potential methanol oxidation. The stability of the Pt particles deposited on diamond has been compared to a Pt impregnated sp2 carbon cloth electrode reported (45) Tilquin, J. Y.; Cote, R.; Guay, D.; Dodelet, J. P.; Denes, G. J. Power Sources 1996, 61, 193.
Langmuir 2009, 25(17), 10329–10336
Figure 8. Chronoamperometry at 400 mV vs Ag/AgCl at 24 and 62 °C of Pt particles deposited on boron-doped diamond films by cyclic voltammetry on 0.5 M CH3OH/0.5 M H2SO4 sample A at 100 mV/s ( 3 3 3 ), sample B at 250 mV/s (;), and sample C at 500 mV/s (---).
by Swain et al. The carbon cloth was observed to fail during the electrolysis, in contrast to the behavior of the Pt particles, which showed no morphological or microstuctural damage. More importantly, no loss of catalytic activity for hydrogen evolution or oxygen reduction was observed after harsh electrolysis.14 The surface characterization of the particles was also performed after methanol electrooxidation. SEMs are shown in Figure 9. Interestingly, for all samples, the only particles that remained attached to the diamond film were the ones deposited on the more conductive facets. This observation shows that not only are the Pt particles preferentially deposited on the conductive (111) facet but also their stability is superior. A possible explanation for this is that each facet of diamond, after oxidation, has different functional groups that participate in the attachment of the platinum particles to the BDD surface, as well as in the nucleation for particle growth. The amount of oxygen available to DOI: 10.1021/la8035055
10335
Article
Gonz alez-Gonz alez et al.
powder particles have been shown to exhibit a preponderance of (111) facets.46 On the basis of the present results, such materials would be anticipated to provide excellent stability for deposited platinum nanoparticles.
Conclusions
Figure 9. Scanning electron microscopy images of polycrystalline free-standing BDD films after Pt deposition by cyclic voltammetry in a 1 mM K2PtCl6/0.5 M H2SO4 solution by cyclic voltammetry at different scan rates (A) 100 mV/s, (B) 250 mV/s, and (C) 500 mV/s after methanol oxidation measurements.
coordinate the platinum atom to be reduced is four, compared with two and three for facets (110) and (100), respectively. Figure 9 shows especially sample A in which the platinum particles agglomerated on a grain boundary of the film, leaving a mostly clean facet of the diamond adjacent to a completely decorated facet, confirming that the stability of the Pt particles on the less conductive facet was poor. This observation confirms that the functional groups present on the oxidized surface of the diamond play an important role in the attachment of the particles to the diamond. It is interesting to highlight an aspect that could have relevance for fuel cell applications, which is that detonation diamond (46) La Torre-Riveros, L.; Tryk, D. A.; Cabrera, C. R. Rev. Adv. Mater. Sci. 2005, 10, 256.
10336 DOI: 10.1021/la8035055
The main purpose of this study is to learn about the chemical properties of the attachment of the platinum particles on the diamond surface because we are presenting it as a possible electrocatalytic support material for fuel cell applications. We have shown that not only the conductivity of the diamond facets plays a role in the kinetics of the Pt particle growth but also the functionalization of the diamond surface plays a role on the nucleation sites and the long-term stability of the deposited Pt nanoparticles. The present work showed that Pt nanoparticles could be deposited on freestanding BDD films by means of cyclic voltammetry, by performing a fast potential sweep in a platinum saltcontaining electrolytic solution. The films were characterized prior to deposition with XRD, Raman, and SECM. The latter showed that, when a potential is applied, the diamond facets exhibit different currents. Thus, there are differences in conductivity among the different facets of the diamond surface. EDS, cyclic voltammetry, and XPS verified the platinum particle chemical composition. From high resolution XPS, platinum particles were mostly in the reduced state, but some of the platinum was oxidized to the platinum dioxide state. Functional groups on the diamond surface were consistent with those expected for the respective facets. SEM images showed that the particle sizes and locations are affected by the sweep rate. The particle deposition depended on the conductivity of the facet and the rate of the nucleation process. Surface characterization, after methanol oxidation, showed that the stability of the particles depended on the facet on which they were deposited. Only the particles deposited on the (111) facet remained attached to the BDD film for all of the samples. Functional groups on the diamond surface, e.g., -OH on the (111) facet, play an important role in the stability of the attachment of the particles to the diamond surface. After methanol oxidation, the Pt particles deposited on other facets appeared to have a loose adhesion and tended to agglomerate at grain boundaries. We showed by high resolution XPS that the -OH functionalization on the diamond surface is involved on the platinum particle nucleation and stability of these on the surface, contrasting with gold particles, which do not adhere to oxidized diamond surfaces, as shown by Kondo et al.18 For the use of diamond as an electrocatalyst support, detonated diamond powder, with a predominance of (111) facets, appears most promising because the adhesion of the particles is expected to be strong on these facets. BDD is thus a promising material as a support for electrocatalyst particles for PEMFCs because of its chemical stability and ability to mitigate the agglomeration of platinum nanoparticles. Acknowledgment. This work was supported in part by NASAURC grant numbers NCC3-1034 and NNX08BA48A and NASA-EPSCoR grant number NNX08AB12A. I. GonzalezGonzalez acknowledges support from a NASA training grant NNG05GG78H (PR Space Grant) and an NACME Alfred P. Sloan Foundation Scholarship. We also acknowledge Mr. Gabriel Cruz of Hewlett-Packard-Puerto Rico for all of his help with the SEM/EDS. We would like to also acknowledge the contribution of the UPR Materials Characterization Center. Langmuir 2009, 25(17), 10329–10336
